Instructor:
   Prof. Sinisa Todorovic
   sinisa at eecs oregonstate edu
   2107 Kelley Engineering Center

Classes:
   TR 8:30-9:50am, Milam Hall 234

Office hours:
   W 4-5pm, or by appointment

Late homework assignments will not be accepted without prior approval.

Homework can be implemented in any programming language on any machine of your choosing. Collaboration on homework is permitted. However, copying of code and reports is not allowed. Homework submission should include a brief description of the implemented algorithm, the listing of a well-commented source code, and experimental results.

Late project assignments will not be accepted without prior approval.

This work involves developing a system that addresses any vision problem of your choosing, and the preparation of a project proposal, in-class presentation, and a final project report. The project can be implemented in any programming language on any machine of your choosing. You are encouraged to work in teams.

Suggestions for Interesting Course-Project Topics

• Recognizing panoramas, panorama stitching, video panoramas
  
  
• Video summarization
  
  
• Human activity recognition
  
  
• Low-level image segmentation
  
  
• Texture segmentation, texture classification
  
  

Instructions for preparing the project proposal (due 11/04 in class)

Each team should submit a one-page write-up that describes:

• Members of the team
  
  
• Statement of the vision problem to be addressed
  
  
• Overview of the method that will be developed
  
  
• Motivation for proposing this method
  
  

Instructions for preparing the in-class presentation (12/02 and 12/04)

Each team should prepare a 30min long presentation, which usually amounts to 25-30 slides. Talks that happen to take longer than 30min will be stopped, and graded only based on the material presented until the interruption. The presentation should include the following:

• Problem statement: What vision problem have you addressed
  
  
• Brief review of prior work addressing the same problem
  
  
• Brief review of related prior work that addresses the same problem using an approach similar to yours
  
  
• Overview of your approach
  
  
• Motivation for using your approach to address the problem
  
  
• List of contributions: Is there anything novel that you have proposed, and, if yes, how this novelty advances the state of the art
  
  
• Detailed presentation of each step of your approach
  
  
• Experimental results done, and overview of additional experiments that you plan to carry out until the final project report is due
  
  
• Concluding remarks: Have you successfully addressed the problem; where and why your approach fails; and what would be your suggestions for the future work
  
  

Instructions for preparing the final project report (due 12/08)

The final project report should be written as a standard, eight-page, double-column, vision-conference paper (e.g., CVPR, ICCV paper). The formatting instructions and example papers are given in cvprkit.zip. Reports that are longer than 8 pages will be graded only based on the material presented in the first 8 pages. While a high quality of this writeup will be expected, it need not be publishable. The final project report should include the following:

• Title, and names of the team members
  
  
• ABSTRACT -- motivation, what is done, and main contributions
  
  
• INTRODUCTION
  
  
  • Problem statement: What vision problem have you addressed
    
    
  • Overview of your approach
    
    
  • Motivation for using your approach to address the problem
    
    
  • List of contributions: Is there anything novel that you have proposed, and, if yes, how this novelty advances the state of the art
    
    
• RELATIONSHIPS TO PRIOR WORK
  
  
  • Brief review of prior work addressing the same problem
    
    
  • Brief review of related prior work that addresses the same problem using an approach similar to yours
    
    
  • Positive aspects of prior work that your approach draws from
    
    
  • Negative aspects of prior work that you have attempted to address
    
    
• DETAILS OF EACH STEP OF YOUR APPROACH
  
  
• EXPERIMENTAL EVALUATION
  
  
  • Qualitative evaluation (if applicable)
    
    
  • Quantitative evaluation
    
    
• CONCLUSION
  
  
  • Main contributions
    
    
  • Main conclusions regarding your experiments
    
    
  • Have you successfully addressed the problem? If yes, what are the main reasons for success?
    
    
  • Where and why does your approach fail? How can these problems be addressed in the future?
    
    
  • General suggestions for the future work
    
    
• ACKNOWLEDGMENT
  
  
• REFERENCES